17:29 <landscape15[m]> j`ey: which options should I enable in defconfig to get Wi-Fi support? I’ve compiled it with `make ARCH=arm64 CROSS_COMPILE=aarch-linux-gnu- defconfig`

17:30 <landscape15[m]> * with ` make ARCH=arm64, * CROSS_COMPILE=aarch-linux-gnu- defconfig `

17:31 <landscape15[m]> s/`//

17:31 <j`ey> at least BRCMFMAC BRCMFMAC_PCIE

17:32 <Glanzmann> landscape15[m]: Best is to use mps config, you find it here: https://tg.st/u/0weE that is tested.

17:33 <Glanzmann> landscape15[m]: Works for mps on the pro and for me on the air.

17:34 <j`ey> and you need to run the firmware script from the asahi-installer repo too

17:35 <landscape15[m]> Glanzmann: I already tried your .config and for me works perfectly.

17:35 <landscape15[m]> I just want to try myself

17:35 <Glanzmann> landscape15[m]: That is becasue it is a copy of mps config. I tried to get the config right by myself but failed.

17:35 <Glanzmann> landscape15[m]: Have fun.

17:36 <Glanzmann> landscape15[m]: I had a config that were missing the options necessary for wpa2. And than I tried to adopt the debian default config and failed. Than I finally used mps config for now.

17:37 <j`ey> probably want BRCM_TRACING and BRCMDBG too

17:38 <Glanzmann> j`ey: BRCMFMAC BRCMFMAC_PCIE this should be enough for the wifi device, but if you have wpa2 you need a lot of additional options. That is were I failed ...

17:38 <landscape15[m]> j`ey: ok thanks. Do I find these options in arm64 defconfig?

17:38 <mps> Glanzmann: right, you need some kernel crypto options/drivers

17:39 <mps> and RFKILL

17:39 <landscape15[m]> Glanzmann: yeah at first I’ll disable wps2 on my modem

17:39 <landscape15[m]> s/wps2/wpa2/

17:39 <j`ey> landscape15[m]: open the config and see if theyre on!

19:01 <Glanzmann> landscape15[m]: When I tried, I used 'make menuconfig'. Btw. Once you have an m1 machine running, you no longer need to crosscompile. That makes it easier to install modules for example.

19:02 <Glanzmann> Glanzmann: I don't know if you go with debian, but if you go that would install build dependencies for the kernel: sudo apt-get install build-essential bc kmod cpio flex cpio libncurses5-dev libelf-dev:native bison libssl-dev

19:03 <Glanzmann> And yes, you need cpio only once, just fixed it in my 'knowhow' file.

19:03 <tpw_rules> someone else should try nixos. i need to get wifi working properly with it though

19:04 <tpw_rules> allegedly keyboard support is in uboot now

19:05 <Glanzmann> Sorry, I'm a Debian fanboy.

19:05 <Glanzmann> tpw_rules: Nice progress on the keyboard, I saw it, but decided to only roll with it once marcan has finished the official installer.

19:06 <Glanzmann> I'm also looking on getting hdmi back on the mini.

19:06 <tpw_rules> why? those are orthogonal

19:06 <Glanzmann> tpw_rules: I know, but I have a setup that works for now on the air, so I'm happy.

19:07 <tpw_rules> oh, i erase mine constantly

19:09 <tpw_rules> i need to find someone unsuspecting to lend me their m1 max

19:12 <Glanzmann> tpw_rules: Last time I erased my mini, I did not get it back. :-)

19:12 <tpw_rules> i mean i don't fully erase it

19:14 <Glanzmann> tpw_rules: Normally I would not do that either but I wanted to make a quickstart video using the mini because there I could capture HDMI. And than it was gone ...

19:15 <tpw_rules> sad

19:15 <tpw_rules> what have you done in pursuit of the goal of getting hdmi back btw?

19:16 <tpw_rules> because hdmi still works for me but sucks completely

19:16 <tpw_rules> i was toying around with the idea of horrendously hacking a driver into uboot to bring things up before linux but the only source of info i found is that dcp.py script in the m1n1 repo

19:18 <sven> DCP has a iboot protocol that should be much simpler than the real one. It’s just not documented anywhere and kinda annoying to reverse engineer since we can’t observe the code that uses it

19:19 <tpw_rules> is it not possible to get an unencrypted iboot image?

19:19 <sven> Not for the M1

19:19 <tpw_rules> (and is that protocol what marcan was referring to when he was talking about "fixing" this stuff? or was he gonna do the real one somehow?)

19:20 <sven> pretty sure that’s the one he was talking about

19:20 <sven> it’s not impossible to reverse engineer since you can e.g. crash DCP very early and see iboot‘s last messages and also look at the DCP firmware itself

19:20 <sven> it‘s much more annoying though

19:48 <Glanzmann> tpw_rules: I just tried to downgrade the mini, but was unable to install m1n1 or get a picture back. I also tried dcp.py. What would probably work, but I have no tried is resetup the mini with Macos 11 and alyssa also has a tree with a dcp driver that worked for me (after she modified it to work on my monitor) that someone could checkout. It also has HDMI hotplug support, IIRC.

19:48 <mps> u-boot works fine from external usb disk but not on internal nvme drive, say it can't find partition with grub, at least in my try. maybe it works fine for someone else even on nvme

19:50 <Glanzmann> mps: I see.

20:00 <mps> this is for more tests when we get SMC

21:11 <tpw_rules> mps: what uboot and what system?

21:12 <tpw_rules> Glanzmann: which branch?

21:17 <mps> tpw_rules: this one https://github.com/kettenis/u-boot/commits/apple-m1-m1n1-nvme

21:18 <tpw_rules> what machine?

21:18 <tpw_rules> is the partition grub is installed on labeled as an EFI system partition?

21:18 <mps> MB PRO, j293 model

21:19 <tpw_rules> so that's the regular M1?

21:20 <mps> efi partition is correct and properly labeled and bootefi64 is there, grub is on nvme0n1p5 partition

21:20 <mps> tpw_rules: yes, mb pro 2020 year model

21:21 <tpw_rules> can you post a log?

21:21 <tpw_rules> the hypervisor would be the best way to get it

21:22 <mps> tpw_rules: how to 'catch' log when machine is not booted, only can make photo

21:22 <tpw_rules> the hypervisor, uboot will dump it over serial

21:23 <mps> ok, will look tomorrow if I can get it

21:23 <tpw_rules> also "bootefi64"?

21:23 <tpw_rules> you need to install grub with the --removable flag even on the internal device

21:24 <tpw_rules> uboot doesn't have efi vars, so it will only look in \efi\boot\bootaa64.efi on the ESP

21:24 <tpw_rules> i think

21:24 <mps> tpw_rules: yes, that file is there

21:26 <mps> it in root of part as EFI/boot/bootaa64.efi

21:27 <mps> and it works fine on usb disk

21:33 <Glanzmann> tpw_rules: IIRC it was 20211022 in alyssas tree.

21:33 <tpw_rules> where is her tree?

21:33 <j`ey> https://github.com/mu-one/linux

21:33 <Glanzmann> tpw_rules: https://github.com/mu-one/linux

21:34 <tpw_rules> hm, might have to pull out some patches

21:36 <Glanzmann> tpw_rules: Alyssa tried recently to get her dcp driver running but gave up after an hour or so.

21:36 <tpw_rules> hm?

21:36 <tpw_rules> on a later kernel or something? did you not just say it works

21:37 <Glanzmann> tpw_rules: But with this dcp driver I was able to change resolution and rotate the screen by 90 degrees. Also cosnole blanking was working and you could wake up the display again.

21:37 <j`ey> tpw_rules: she was trying to rebase it

21:37 <Glanzmann> tpw_rules: Yes, she tried on top of Linus or Asahi Head, I don't remember.

21:38 <Glanzmann> She was frustrated and gave up after a lot of patches were not working.

21:39 <tpw_rules> i guess it's not just that one DCP commit

21:44 <tpw_rules> Glanzmann: how does one change the resolution? xrandr i guess?

22:40 <tpw_rules> mps: does the keyboard work for you?

22:48 <mps> tpw_rules: yes

22:48 <mps> sorry for late answer, I was afk

22:49 <tpw_rules> that's good to know, i'll update my docs

22:49 <mps> twoscomplement0b[m]: what docs? url?

22:50 <mps> btw, arrow keys doesn't work, must use C-P and C-N

22:50 <mps> twoscomplement0b[m]: sorry intended to tpw_rules

22:51 <tpw_rules> https://github.com/tpwrules/nixos-m1/blob/main/docs/uefi-standalone.md this one

22:51 <mps> tpw_rules: thanks

22:52 <tpw_rules> is there any chance you can test the kernel config here: https://github.com/tpwrules/nixos-m1/blob/main/nix/kernel/config to see if the keyboard works wiht it

22:52 <tpw_rules> i have options at the bottom which i think are enough but it would be good to be sure

22:57 <mps> tpw_rules: what you mean, don't understand. I have kernel which works fine, I can paste config if you want to compare with one above config

22:58 <tpw_rules> i'm not sure if mine has all the relevant options

22:58 <tpw_rules> and i don't have a machine with a builtin keyboard to test

22:58 <mps> ah

22:59 <mps> tpw_rules: here is my latest config https://tpaste.us/xnqD

23:00 <mps> probably some things not needed but I don't have time nowadays to test all of them, maybe I will do when SMC is pushed and got u-boot working on nvme

23:00 <mps> and then testing kernels will be a lot easier, I think

23:00 <tpw_rules> u-boot works great on nvme for me on mac mini. i might be able to help you troubleshoot it later

23:01 <mps> as I wrote above I will try tomorrow again if something doesn't 'block' me

23:02 <kettenis> mps: efi/boot/bootaa.efi needs to be on msdos/vfat partition

23:02 <kettenis> if you put it on your linux root partition, it won't work

23:02 <mps> kettenis: it is

23:03 <mps> I created it FS with 'mkfs.fat -F 32 ...'

23:04 <tpw_rules> does it?

23:05 <mps> but iirc u-boot will find files even on ext4

23:05 <kettenis> the GPT partition type also has to be EFI Sys

23:05 <mps> kettenis: it is also

23:06 <mps> EFI System

23:06 <mps> in fdisk